基于随机森林的大型公共建筑能耗混合预测模型  

A Hybrid Prediction Model for Energy Consumption of Large Public Buildings Based on Random Forest

作　　者：于军琪[1] 虎群 赵安军[2] 高之坤 成浩 张娜[2] YU Jun-qil;HU Qunl;ZHAO An-jun;GAO Zhi-kun;CHENG Hao;ZHANG Na(School of Information and Control Engineering,Xi'an University of Architecture and Technology,Xi'an 710055,China;School of Building Services Science and Engineering,Xi'an University of Architecture and Technology,Xi'an 710055,China;China Northwest Architecture Design and Research Institute Co.,Ltd.,Xi'an 710015,China)

机构地区：[1]西安建筑科技大学信息与控制工程学院,陕西西安710055 [2]西安建筑科技大学建筑设备科学与工程学院,陕西西安710055 [3]中国建筑西北设计研究院有限公司,陕西西安710015

出　　处：《控制工程》2023年第4期597-604,共8页Control Engineering of China

基　　金：西安机场三期扩建工程能源站办公系统智能控制与咨询项目(20210103)。

摘　　要：针对公共建筑能耗预测模型中影响变量相关性低、冗余性高的问题,提出了基于二氧化碳浓度的公共建筑人员流动率间接测量方法,以提高模型的预测精度,并提出了一种大型公共建筑能耗混合预测模型。首先利用LASSO变量选择算法筛选出与公共建筑能耗相关性高的影响因素,再引入改进的并行排序蚁群优化算法对随机森林预测模型的参数进行优化,进一步提高预测性能。最后,以西安某公共建筑监测数据为例进行预测分析。结果表明,人员流动率对公共建筑能耗预测有着重要的影响,所提模型的泛化能力强、预测精度高,可以为公共建筑节能优化提供有效的数据支撑。Aiming at the problems of low correlation and high redundancy of influencing variables in the energy consumption prediction model of public buildings,an indirect measurement method of personnel turnover rate in public buildings based on carbon dioxide concentration is proposed to improve the accuracy of the prediction model,and a hybrid prediction model is also proposed to predict the energy consumption of large public buildings.Firstly,the LASSO variable selection algorithm is used to extract the main influence factors which are highly correlated with energy consumption.Then,the improved parallel-ranking ant colony optimization algorithm is introduced to optimize the parameters of random forest model and further enhance the prediction performance.Finally,the energy consumption monitoring data of a public building in Xi'an province is taken as an example for prediction analysis.The analysis results show that the personnel turnover rate has an important influence on the energy consumption prediction of public buildings.The proposed model has good generalization ability and high prediction accuracy,which can provide effective data support for energy conservation optimization of public buildings.

关 键 词：公共建筑能耗 人员流动率 LASSO变量选择 改进的并行排序蚁群优化算法 

分 类 号：TU831[建筑科学—供热、供燃气、通风及空调工程]